WELDING PROCESS
In the initial welding phase, i.e. during the initial plunging stage, the rotational speed was set to 750 rpm, Fig. 2a. As soon as the material around the probe is plasticized, the rotational speed is set to the target values, i.e. to 500 rpm and 400 rpm for Welds I and II, respectively. At this point, the pipe starts to rotate at a welding speed of 2 mm/s. The process is performed force-controlled, with forces of 50 kN and 45 kN for Weld I and Weld II, respectively. Since the process parameters in Weld II led to lower temperatures, Fig. 2b, the higher material resistance can be noticed from the machine response in terms of a slightly higher torque compared to that of the Weld I.
Microstructural Characterization
The microstructural features of the BM are shown in Fig. 3. The X65 steel is composed of a large amount of fine and acicular ferrite (AF) with some quasi-polygonal ferrite (QPF) and polygonal ferrite (PF) (Fig. 3a), while the CRA layer (Inconel 625) presents a microstructure of elongated austenitic grains and M(C, N) carbonitrides. In addition, it is possible to see carbides along grain boundaries, which typically are identified as M6C and/or M23C6. (Fig. 3b). Due to the hot roll-bonding process used to produce the clad pipes, involving high deformation at elevated temperature, substitutional inter-diffusion of Ni and Cr to the steel side and Fe to the CRA side occurred, besides interstitial diffusion of C from steel to the CRA. That results in an austenitic steel in the interface (see in Fig. 3c) due to the strong stabilizing effect of Ni [16], [18]. Figure 3d) shows the results of the EDS line scan along the interface of the pipe, showing a clear transition from steel (Fe) to Inconel (Ni and Cr). Microhardness tests of the BM led to 210 HV, 249 HV and 318 HV for steel, interface, and Inconel 625, respectively.
The two welded joints are shown in Fig. 4. Weld I (Fig. 4a) resulted in a top surface with more flash and larger heat-affected zone (HAZ), Fig. 4b, when compared with Weld II (Fig. 4c, d). Latter is expected due to the higher energy input, see Table 2, and the resulting higher temperature, Fig. 2b). The temperatures in the tool reached a maximum of 1020°C for Weld I and 916.5°C for Weld II.
The joint microstructure can be divided into three core parts: steel side, Ni-based alloy 625 side and interface. On the steel side, Fig. 5a, the SZX65 and hard zone (HZX65) are observed and similarly to other studies [19]–[21], the HAZX65 can be subdivided into three sub regions with different microstructures: the outer HAZ (OHAZX65), middle HAZ (MHAZX65), and the inner HAZ (IHAZX65). The SZX65, Fig. 5b, comprises AF and coarse granular bainite (GB). Additionally, since HZX65 may have experienced the highest deformation, peak temperatures and cooling rates during FSW [22], a lath bainite microstructure (LB) is observed, Fig. 5c. The OHAZX65, Fig. 5d, underwent recrystallization, forming a more refined equiaxed PF compared to the BMX65. In the MHAZX65, the resulting microstructure is composed of QPF, PF and AF, Fig. 5e. The IHAZX65, Fig. 5f, exhibits a mixed ferrite and bainite microstructure with finer prior austenite grain size and more PF than found in the SZX65.
Figure 6 presents the microstructural zones on the Inconel side. Two sub regions can be identified in SZ625, i.e. SZ1625 and SZ2625.The FSW process led to a significant grain refinement in the bottom of SZ1625, Fig. 6a. In SZ2625 coarser and equiaxial austenitic grains are found, Fig. 6b, which suggests that the material was exposed to higher thermal cycles in relation to SZ1625, causing grain growth. The TMAZ625 presents a deformed microstructure following the probe flow pattern, Fig. 6c.
The interface of the welds is detailed in Fig. 7. As Inconel 625 and API X65 steel have the same crystal structure (FCC) at the joining temperature during FSW, as well as, similar melting points and comparable flow stresses [23] these characteristics allow the Ni alloy to flow around the probe and to drain into the steel, forming alternating bands of materials in the SZ, which is consistent with findings from Rodriguez [2]. This is also in agreement with other studies at the interface of dissimilar FSW welds [24]. As can be seen in Fig. 7b-c for both welds, on the advancing side (AS) and on the retreating side (RS), asymmetric Inconel 625 hooks were formed in API X65 steel. The hooks height are around 1.38 mm (0.86 mm) on the AS and 0.84 mm (1.71 mm) on the RS for Weld I (Weld II). The shape of the alternating bands, Fig. 7c-d, might be related to the process parameters and heat input achieved. A higher heat input seems to contribute to a more homogeneous mixture of the materials.
Due to the temperatures reached during FSW, interdiffusion between the dissimilar materials may occur. To evaluate the diffusion of Ni, Cr and Fe in the SZ area, the chemical composition of three different locations next to a Fe-Ni mixture was analyzed, Fig. 7f). The EDS 1 showed the chemical composition of a steel with a high content of chromium (96.66% Fe, 1.55% Cr, 0.93% Mo), EDS 2 showed the chemical composition of Inconel 625 (4.26% Fe, 64.40% Ni, 21.05% Cr, 6.83% Mo) and EDS 3 showed the chemical composition of a layer with high diffusion of elements (52.06%Fe, 23.84% Ni, 9.47% Cr and 5.82% Mo). Precipitate transformation compound, on the other hand, was not found in the welds in a SEM investigation. According to results in the literature, carbide precipitations are unlikely to occur during the FSW process, as a prolonged exposure at elevated temperatures might be required. [2], [5], [25].
MICROHARDNESS
Vickers microhardness results for Weld I and Weld II show differences in all zones of the Inconel 625, Fig. 8. In the BM near to the weld, values ranged between 310 HV and 360 HV, for Weld I and Weld II, respectively. On the RS, where coarse recrystallized Inconel 625 grains were found, Fig. 7c), the maximum value for Weld I was 284 HV and 330 HV for Weld II. The values in the SZ are significantly lower due to the mixing between Inconel 625 and steel, i.e. for Weld I the values ranged from 218 HV to 303 HV and for Weld II even between 187 HV to 344 HV. The complex banded patterns, see Fig. 7, contributed to the high hardness variations, as the base materials of steel and Inconel 625 have different metallurgical and mechanical properties. On the steel side, the minimum hardness is found in the HAZ, varying between 172 HV and 180 HV, which is in agreement with the literature [14], [22]. Comparing the two welds, the hardness increases slightly with decreasing energy input while the width of the HAZ decreases. Within the SZ, a slight increase in microhardness was noted, ranging between 200 HV and 250 HV for both welds. This slight increase in microhardness in the SZ is expected due to the severe deformation at elevated temperatures and high cooling rate, resulting in a greater amount of bainite in the steel [19], [26]. In the SZ, higher strain rates and associated plastic deformation caused higher hardness zones on the AS in comparison to that of the RS. The BM interface region had about 245 HV.